The reptile market "consumes" a large number of day old male chicks. I suspect some are also redirected to dogfood / catfood.
You don't have to grow *every* male out to maturity in order to cull it. In this case cull means removing from the breeding program. If you sell a dog that doesn't fit your breeding program as a pet, you've culled it. That it gets to live out is life as a human companion is awesome and irrelevant.
Every good breeding program should have multiple layers of selection. If your breed spec says "must have feathered feet", you can cull males and females on day one that don't have feathered feet.
Autosexing in birds (which is different than sex linked) is usually? Always? Based on the dilution factor. I think. Maybe there's another technique. The males get two copies and are white(ish) and the females get a single copy and are... some other color. Sex linked is usually built on the barred gene, and parents of the appropriate color & pattern need to be selected for it to work with a single generation. Those offspring won't themselves produce the same results. Autosexing works in each generation.
People like autosexing birds, for all the reasons listed above, and more. It's fun to look out at your pilgrim geese and know by color which one is the gander. For some people it is just fun to have color coded birds that are all the same breed. That's it. That's why autosexing is a thing. I'm not aware of any commercial variety that is autosexing, because sex linked is simpler and more in line with the style of hybrid breeding that goes into commercial varieties.
So as an example, if you were breeding a clean footed, walnut combed, early maturing breed you'd cull at least three times.
Day one anything that has feathered feet, isn't autosexing, and obviously has a large single comb gets cut from the breeding program. Later as it becomes obvious that something has a pea or a rose comb, then that gets culled. And the ln finally anything that isn't crowing or laying at 7 or 6 or 5 months, depending on where you're at, gets cut.
And no, I can't recall the genetics on walnut combs. If that's the one that's a partially dominant mic of rose and pea or such, pretend that I bothered to look it up and use the correct example. Thanks.